基于场论的本体中概念相似度求解方法

Concept similarity measuring method for ontology based on field theory

针对目前本体中概念相似度求解方法孤立考虑概念信息量或概念关系的问题,提出基于场论的概念相似度求解方法。该方法首先明确概念的语义关系是由概念包含的信息决定且隐性的概念关系通过显性的概念关系推出,并将概念包含的信息划分为共性信息与特性信息两类。结合场论的原理和方法,将本体映射为与物理场具有相同特性的语义场,将概念的共性信息和特性信息分别转化为语义引力元和语义距离,采用引力场模型计算概念的相似度。结合坦克装甲车辆和舰船领域知识检索需求,建立基于本体的语义检索系统,并通过与五种典型的相似度求解方法进行对比分析,验证了该方法的正确性和有效性。

For the problem that current methods for concept similarity in ontology considered the concept information or concept relationship in isolation,a field theory-based concept similarity measuring method was proposed.The method defined that the semantic relationship between concepts was decided by the information that concepts contained and the implicit concept relationship was induced from the explicit concept relationship.The information of concept was divided into common information and special information.Based on above conclusion and field theory,ontology was mapped into a semantic field which had the similar characters with physical field.Common information was taken as semantic gravity elements and special information was taken as semantic distance.The gravity field model was used to calculate the concept similarity.To meet the requirement of armored vehicle and ship＇s design,the ontology-base semantic search software was developed.The effectiveness and correctness of the proposed method were also validated through comparing with other five kinds of similarity measuring methods.